On 03/10/2014 14:32, Bert Verhees wrote: > >> Thanks for getting involved in this important discussion. I am on a >> plane and don't have ooportunity for a full answer but it is not as >> simple as u have suggested. Amongst other things You have to consider >> what happens to archetype numbering when the archetype has been >> published and goes back into review, at which point it is in an >> unstable state. More > > > That is why you often see in addition to Semver-conventions some terms > added. > My bash-version for example is: GNU bash, version 4.3.11(1)-release, > My Linux version is: 3.13.0-36-generic > > Most artefact producers have the need to communicate more then only > the version-number. The discussion is if the version-number is the > place to communicate this. > Software too can become stable, unstable, unsafe, release, beta, > without anything changed to the bits and bytes. > > But suppose it is: > > Suppose you have version 1.1.3-release, and then you have > 1.1.3-unstable, which is later? You cannot tell.
the latest rules that Ian and Sebastian have worked out would mean that if there was 1.1.3 released, then the next version of that archetype if it goes back to development is 1.1.4-unstable, i.e. at least the patch number has to be incremented. It could be that something more is incremented say 1.2.0-unstable. So I think this is pretty logical (and also pretty much the way we do it in software). - thomas
